2005 customers to upgrade, which begs the question: Which version should they upgrade to?
Joseph D’Antoni, principal consultant at Denny Cherry & Associates Consulting, recommended upgrading to the latest SQL Server version that supports the third-party applications a company is running. He said there are big changes in between each of the versions of SQL Server, adding that SQL Server 2014 is particularly notable for the addition of the cardinality estimator. According to D’Antoni, the cardinality estimator can “somewhat drastically” change query performance for some types of data. However, the testing process is the same for all of the versions, and the same challenges — testing, time and licensing — confront any upgrade. “You’re going to have a long testing process anyway. You might as well try to get the latest version, with the longest amount of support.”
“If it were me, right now, contending with 2005, I would go to 2014,” said Robert Sheldon, a freelance writer and technology consultant. “It’s solid, with lots of good features. There would be no reason to go with 2012, unless there were some specific licensing circumstances that were a factor.” Denny Cherry, founder and principal consultant at Denny Cherry & Associates Consulting, recommended upgrading to SQL Server 2012 at the earliest, if not 2014, because “at least they won’t have to worry about upgrading again anytime soon.”
Although SQL Server 2014 is the most current SQL Server version, SQL Server 2016 is in community technology preview. Sheldon said he doesn’t see upgrading to SQL Server 2016 as a good strategy. “Those who want to upgrade to SQL Server 2016 face a bit of a dilemma, because it is still in preview, and I have not yet heard of a concrete release date,” he said. “An organization could use CTP 3.0 to plan its upgrade strategy, but I doubt that is something I would choose to do.”
D’Antoni considered the possibility of waiting until the release of SQL Server 2016 to upgrade. “If they identify a feature that’s compelling, maybe they should wait for 2016,” he said. He added that “2016 is mature enough to roll,” and the only real problem is that it is currently unlicensed.
“If they’re already out of support and planning on moving to 2016, it could be worth waiting the few months,” Cherry said. Furthermore, Cherry said, waiting for SQL Server 2016 could save an organization from having to go through a second upgrade in the future.
Cherry added that, for everyone not waiting for SQL Server 2016, “If they haven’t started the project yet, they should get that project started quickly.” D’Antoni had an even more advanced timetable. He said a company “probably should have started already.” He added, “It’s the testing process that takes a lot of time. The upgrade process is fairly straightforward. Testing the application to make sure it works should have started fairly early.” Ideally, D’Antoni said, by this point, organizations should have done some initial application testing and be planning their migration.
A number of Cherry’s clients, ranging from small businesses to large enterprises, are upgrading because of the approaching SQL Server 2005 end of life. He described SQL Server 2005 end of life as affecting “every size, every vertical.” D’Antoni predicted the small organizations and the largest enterprises will be the hardest hit. The small corporations, he said, are likely to be using SQL Server 2005, because they lack the resources and IT personnel for an easy upgrade. The large enterprises, on the other hand, have so many systems that upgrades become difficult.
D’Antoni explained that, while it is possible to migrate to an Azure SQL database in the cloud instead of upgrading to a more advanced on-premises version of SQL Server, he doesn’t expect to see much of that — not because of difficulties with the product, but because of company culture. Companies who use the cloud, he said, are “more forward-thinking. If you’re still running 2005, you tend to be less on top of things like that.”